AQUATIC MICROBIAL ECOLOGY Vol. 38: 181–190, 2005 Published February 9 Aquat Microb Ecol

Sediment properties and bacterial community in burrows of the ghost shrimp Pestarella tyrrhena (: Thalassinidea)

Sokratis Papaspyrou1,*, Trine Gregersen2, Raymond P. Cox2, Maria Thessalou-Legaki1, Erik Kristensen3

1Department of Zoology–Marine Biology, Faculty of Biology, University of Athens, Panepistimiopolis, 157 84 Athens, Greece 2Department of Biochemistry and Molecular Biology, and 3Institute of Biology, University of Southern Denmark, 5230 Odense M, Denmark

ABSTRACT: Chemical properties of burrow wall sediment from burrows of the thalassinidean shrimp Pestarella (=) tyrrhena located at Vravrona Bay (Aegean Sea, Greece) were studied and found to be very different from the sediment surface and ambient anoxic sediment. P. tyrrhena burrow walls had significantly higher amounts of silt and clay, while total organic carbon (TOC) was up to 6 times higher than in surrounding sediment. Chlorophyll a (chl a) accounted for a small frac- tion of TOC and showed similar values in burrow walls and surface sediment, whereas the low chl a: chl a + phaeopigment ratio indicated the presence of more fresh material in the latter. Biopolymers (carbohydrates, proteins and lipids) were 4 to 11 times higher in burrow walls than in the surround- ing sediment, accounting for 47% of TOC. The low protein:carbohydrate ratio indicated that the high TOC in the burrow walls was caused by the presence of aged detritus of low nutritional quality, such as seagrass detritus. The distinct conditions along the burrow wall also affected the bacterial community and resulted in a 10-fold increase of bacterial abundance. Molecular fingerprints of the bacterial communities showed that the bacterial composition of the burrow wall was more similar to the ambient anoxic sediment and showed less seasonal change than the sediment surface. These results suggest that burrow walls have distinct properties and should not be considered merely as a simple extension of the sediment surface.

KEY WORDS: Bioturbation · Organic matter · Seagrass detritus · Bacteria · PCR-DGGE · Pestarella tyrrhena

Resale or republication not permitted without written consent of the publisher

INTRODUCTION plex structures (Griffis & Suchanek 1991, Ziebis et al. 1996, Dworschak 2001). Biogenic structures, such as tubes and burrows, are a The burrow environment is usually very distinct and ubiquitous feature of aquatic environments. Tubes and differs from other parts of the sediment (surface and burrows are semi-permanent or permanent structures ambient anoxic sediment). Burrow walls are often rich in that differ in size, appearance and composition accord- organic matter of variable reactivity depending on its ing to the feeding habits, mode of life, mobility, as well origin, chemical composition, structure and age (Aller & as the size of infaunal species (Meadows 1991). Bur- Aller 1986, Reichardt 1988, de Vaugelas & Buscail 1990). rows range in length from a few mm (meiofauna) to The low diffusivity of burrow linings, when present, several meters (large polychaetes and ). reduces transport of solutes between the sediment and They may have vertical or horizontal orientation, be the burrow lumen, while the usually intermittent pattern branched or straight, and have either simple or com- of irrigation promotes very variable oxygen con-

*Email: [email protected] © Inter-Research 2005 · www.int-res.com 182 Aquat Microb Ecol 38: 181–190, 2005

ditions in the burrows (Kristensen 1988, Ziebis et al. 24° 01’ E). The coast is lined by salt marshes and a sea- 1996, Furukawa 2001). The availability of labile organic sonal stream discharges near the head of the bay. The matter combined with steep chemical gradients and tidal amplitude in the area ranges from 10 to 30 cm and narrow redox zonation has a significant impact on the the largest part of the flat is exposed to the air during chemical and biological composition of the burrow spring low tides. The sediment in the sandflat consists environment. Bacterial abundances have been shown to mainly of moderately sorted medium to fine sand with be higher along burrow walls compared to either surface a median particle size of 0.2 mm and organic matter or ambient sediment (Aller & Aller 1986, Branch & content of 2.2%. The macrophyte vegetation is charac- Pringle 1987, Dworschak 2001). In addition, burrow walls terised by scattered occurrence of the seagrass show increased heterotrophic activity by both aerobic Cymodocea nodosa. Dead leaves of another seagrass, and anaerobic bacteria, resulting in increased rates Posidonia oceanica, are continuously transported into of organic matter decomposition (Aller & Aller 1986, the area from deeper waters. Mean salinity in the area Reichardt 1988, Gribsholt et al. 2003). ranges from 20 to 39‰ during the year and water Thalassinidean ghost shrimps have been recognised temperature from 11 to 29°C. in recent years as one of the most effective bioturbating Pestarella tyrrhena is the most abundant large bio- groups of macrofaunal organisms, with significant im- turbator in the intertidal and shallow subtidal area and pacts on the benthic environment (Griffis & Suchanek shows a mean yearly population density of 63 ± 9 ind. 1991, Reise 2002). Pestarella tyrrhena is an important m–2, and a maximum of 144 ind. m–2 (Thessalou-Legaki bioturbator, commonly found in muddy and fine sandy 1987). It occurs in the middle part of the sandflat, intertidal and shallow subtidal coastal sediments, avoiding both the wave-exposed seaward front and where it often creates dense monospecific populations. the muddier inshore salt marsh fringe that is charac- It is a selective deposit feeder that constructs deep and terised by large salinity variations (Thessalou-Legaki complex burrows, constantly digging new branches or 1987). The diversity and abundance of other benthic filling up existing ones (Dworschak 1987). P. tyrrhena macrofauna are generally low (A. Nicolaidou & increases the organic matter content in the sediment M. Thessalou-Legaki unpubl. data). and benthic metabolism by incorporating organic de- Sediment sampling. Sediment samples were col- tritus, such as seagrass debris, in burrow chambers lected at low tide in November 2001, and January, April (Dworschak 1987, 2001, Papaspyrou et al. 2004). In ad- and August 2002. Samples were collected from 3 differ- dition, it has been proposed that P. tyrrhena consoli- ent compartments associated with Pestarella tyrrhena dates the burrow walls with mucus (Dworschak 1983, burrows using a stainless steel spatula: burrow walls, 1998). However, detailed studies have not been made anoxic subsurface sediment adjacent to the burrows on the effect of P. tyrrhena on the distribution and com- (referred to as ‘ambient’) and the sediment surface. position of organic matter associated with the burrow, Undisturbed surface sediment was retrieved to a depth or its effects on the sediment microbial community. of 1 to 2 mm. Burrow wall and ambient sediment were The aims of this study were to investigate the abiotic sampled from 5 to 20 cm below the surface after expos- and biotic burrow environment of Pestarella tyrrhena ing burrows by digging. The burrow wall, defined as and to determine how sediment properties along the layer with a clearly visible difference in colour and burrow walls differ from the surroundings. For this texture from adjacent sediment, corresponding to a purpose, we collected seasonal samples from burrow radial distance of 1 to 2 mm, was recovered first. Sub- walls, surface sediment and anoxic sediment, to deter- sequently, the surrounding ambient sediment was mine particle size distribution, total organic carbon collected at the same depth, but away from the burrow. (TOC) content and relative contributions of major Sediment material was collected from at least 5 differ- organic compounds (proteins, carbohydrates and ent burrows, pooled together and analysed in triplicate, lipids), and phytopigments. In addition we studied the with the exception of bacterial abundance and parti- bacterial communities in burrow walls and the sur- culate iron pool analyses which were performed on rounding sediment by measuring bacterial abundance samples from individual burrows (see later). and obtaining molecular fingerprints of the bacterial Samples for determination of the sediment physico- community based on 16S rRNA gene sequences. chemical characteristics were stored in acid-washed Eppendorf tubes (10% HCl overnight, rinsed with Milli-Q water) and kept on ice until return to the labo- MATERIALS AND METHODS ratory, where they were immediately frozen at –80°C and subsequently freeze-dried. Samples for bacterial Study site. Sample collection was conducted in a DNA extraction were collected in sterile Eppendorf small (0.26 km2) intertidal-shallow subtidal sandflat tubes using a sterile stainless-steel spatula, kept on ice at Vravrona Bay, Aegean Sea, Greece (37° 56’ N, and stored at –80°C until analysis. Papaspyrou et al.: Properties of Pestarella tyrrhena burrows 183

Samples for determination of bacterial abundance Particulate iron. Samples for particulate iron analy- were taken in January and August 2002. Approximately sis were analysed according to the colorimetric method 300 mg sediment from a single burrow was transferred of Lovley & Phillips (1987). Briefly, 300 mg of sediment to a pre-weighed sterile Eppendorf tube containing 1 ml was immediately transferred after collection to 5 ml of filtered (0.2 µm filter) 2% glutaraldehyde in seawater 0.5 M HCl and extracted for 30 min. Subsamples of and stored at 4°C until analysis. Sediment samples for the extractant were added to Ferrozine for Fe(II) deter- measurement of particulate iron pools were taken from mination. Total reactive Fe was determined after sediment cores collected in January 2002 using 8 cm reduction of Fe(III) by hydroxylamine hydrochloride to (internal diameter) plexiglas tubes. The cores were kept Fe(II) in a parallel sample and analysed with Ferrozine. in a water tank with oxygenated water at in situ tem- Reactive Fe(III) was defined as the difference between perature and transported to the laboratory. Cores were total Fe and Fe(II) (Kostka & Luther 1994). split in 2 to expose individual burrows and samples Bacterial counts. Total bacterial abundance was esti- were collected as described earlier. mated by the DAPI (4’6-diamidino-2-phenylindole) Granulometry. Grain size analysis of sediment was counting method (Porter & Feig 1980) as adapted by carried out according to Buchanan (1984). Sediment Andresen & Kristensen (2002). subsamples were also examined for wet density Statistical analysis. Differences between sediment (weight of a known volume) and water content (weight types were tested by the non-parametric Kruskal- loss after drying at 105°C for 6 h). Wallis test. In the case where analyses were performed Organic content. Sediment organic matter (loss-on- on pooled samples seasonal data were used as repli- ignition; LOI) was determined as the weight loss after cates. Bacterial abundance was also tested for seasonal combustion of dried sediment at 520°C for 6 h. TOC differences. Spearman-rank correlation analysis was was measured by the wet oxidation technique (Nelson performed to test for possible relationships between & Sommers 1975). Particulate organic nitrogen (PON) investigated variables. content was analysed using a Carlo Erba EA1108 CHN DNA extraction. Sediment samples for DNA extrac- Elemental Analyser according to Kristensen & Ander- tion were thawed and mixed thoroughly before extrac- sen (1987). tion using enzymatic treatment and phenol extraction Biopolymers. Protein content (PRT) was determined method (Ausubel et al. 1995). Approximately 300 mg of according to Rice (1982) and expressed as bovine material were suspended in 180 µl of lysis buffer serum albumin equivalents. Carbohydrates (CHO) (20 mM Tris, 2 mM EDTA, 1.2% Triton, pH 8.0 with were measured according to Gerchakov & Hatcher HCl) and vortexed for 2 min. Lysozyme was added at a (1972) and expressed as glucose equivalents. Total final concentration of 0.4 µg µl–1 and the suspension lipid content (LPD) was determined after extraction by was incubated at 37°C for 30 min. After the addition of direct elution with chloroform-methanol (Bligh & Dyer 15 µl of Proteinase K solution and 200 µl AL-buffer 1959, Marsh & Weinstein, 1966) and reported as tri- (Qiagen DNA Stool Kit), tubes were incubated at 65°C palmitine equivalents. Proteins, carbohydrates and for 30 min. DNA was purified by phase-separation with lipids were converted to carbon equivalents using con- a buffered solution (pH 7.9) of phenol:chloroform: version factors of 0.49, 0.40, 0.75 g C g–1 respectively, isoamylalcohol (25:24:1). After precipitation with ice- based on the corresponding standard compounds used cold isopropanol overnight at –20°C, the DNA was (Pusceddu et al. 1999). The biopolymeric fraction (BPL) washed in cold 70% ethanol, dried, redissolved in 30 µl was defined as the sum of carbohydrate, protein and of Milli-Q water and stored at –80°C. Four replicate lipid carbon. Combusted sediment (520°C, 6 h) from tubes were used per sediment type and the extracted each sediment type was used as blank samples. DNA was then pooled. Microphytobenthos. Chlorophyll a (chl a) and phaeo- DNA was further purified using a Sepharose 4B col- pigments (phaeo) were determined by the acetone umn (Miller 2001). Briefly, Sepharose 4B was packed extraction method (Parsons et al. 1984). Five ml of 90% by gravity in a 2.5 ml syringe to a final volume of 2 ml. acetone were added to about 0.5 g of sediment and left The column was equilibrated with 4 volumes of high overnight at 5°C. After shaking and centrifugation the salt TE buffer (100 mM NaCl, 10 mM Tris, 1 mM EDTA; concentration was determined spectrophotometrically pH 8.0 with HCl). Crude DNA extract were added to at 665 and 750 nm before and after addition of one the column followed by several additions of 0.2 ml high drop of 10% HCl (Parsons et al. 1984). Chloroplastic salt TE buffer. The eluate was collected in 0.2 ml frac- pigment equivalents (CPE) were calculated as the sum tions. The fractions giving PCR products were pooled of chl a and phaeo concentrations. In order to estimate and frozen at –80°C. the microphytobenthic carbon, chl a concentrations PCR reaction. Each PCR reaction was performed in a were converted to carbon content using a conversion total volume of 20 µl. Reaction mixtures contained final factor of 40 (De Jonge 1980). concentrations of 1 × PCR buffer (Amersham Pharma- 184 Aquat Microb Ecol 38: 181–190, 2005

cia Biotech); BSA 0.5 µg µl–1; deoxynucleoside tri- phosphate solution (200 µM each); universal bacterial primers; GC-341f (5’ CGC CCG CCG CGC CCC GCG CCC GGC CCG CCG CCC CCG CCC CCC TAC GGG AGG CAG CAG 3’) (Muyzer et al. 1993, Over- mann & Tuschak 1997) and 907r (5’ CCG TCA ATT CMT TTG AGT TT 3’) (Muyzer et al. 1993) [M=A/C] (1 pmol µl–1 each); and 1 U of Taq DNA polymerase (Amersham Pharmacia Biotech). DNA templates were prepared by dilution of the purified DNA extracts by 1:100 in Milli-Q water. PCR was performed using the touchdown PCR program of Schäfer & Muyzer (2001). DGGE analysis. PCR products were loaded directly onto the DGGE (Denaturing Gradient Gel Elec- trophoresis) gel. Gels were run at 200 V for 5 h in 1 × Fig. 1. Seasonal variation in total organic carbon content TAE with a 35 to 70% denaturating gradient of (mg [g dw]–1) of surface sediment, burrow wall, and ambient formamide and urea, using the DCode Universal sediment associated with burrows of Pestarella tyrrhena Mutation detection System (Bio Rad). DNA bands were visualised with ethidium bromide. size diameter in the burrow wall (p < 0.05). In addition, The band patterns on the gels were analysed using burrow wall sediment was poorly sorted, having a sig- specially written software. The individual lanes of the nificantly higher silt and clay fraction (19%), compared DGGE gels were digitised and the background sub- to surface and ambient sediment (1%) (p < 0.05), which tracted using the ‘rolling disk’ algorithm (Sternberg both showed medium sorting. 1983). The major peaks were identified manually and a Yearly mean values of organic content (LOI) in the difference matrix was produced using the Dice algo- burrow wall were 3.8 to 4.1 times higher than surface rithm. Clustering was performed using UPGMA (un- and ambient sediment (p < 0.05) (Table 1). TOC showed weighted pair group means with arithmetic averages). the same pattern as LOI. Yearly mean burrow wall TOC was approx. 6 times higher (18.4 mg g–1), than that of surface sediment or ambient sediment (3.11 mg RESULTS g–1 and 3.51 mg g–1 respectively) (p < 0.05) (Fig. 1). TOC showed a significant correlation with all bio- Sediment characteristics polymeric compounds and number of bacteria but not with phytopigments (Table 2). The C:N ratio of the bur- Burrow walls had a smooth texture due to tightly row wall was intermediate to those of surface and compacted muddy sediment and showed a red- ambient sediment (Table 1). brownish colour. Seagrass debris was seen incorpo- rated into the burrow wall. Surface sediment, on the other hand, was light brown while the ambient sedi- Microphytobenthic biomass ment was in most cases dark grey to black. Burrow walls had a lower density than surface and Chl a content showed a higher yearly mean value in ambient sediment, the latter 2 showing similar values the burrow wall (3.70 µg g–1) than surface sediment (Table 1). In all cases the sediment was characterised (2.70 µg g–1) or ambient sediment (0.60 µg g–1), as fine sand, showing, however, a lower median grain- although this was statistically significant only with the latter (p < 0.05) due to the large sea- sonal variations observed. Chl a was Table 1. Sediment characteristics for surface sediment, burrow wall and ambient higher in surface sediment than in sediment from burrows of Pestarella tyrrhena. Md = median grain size diameter, LOI = organic content as loss-on-ignition, C:N = particulate organic carbon to burrow walls in August and lower in particulate organic nitrogen ratio. Values reported are yearly means ± SE (n = 4) January and April, while the 2 sites were similar in the previous November Sediment Wet density Md Silt & clay Sorting LOI C:N (Fig. 2). Mean phaeo content were (g cm–3) (µm) (%) (%) (mol:mol) highest in the burrow wall (8.10 µg g–1) (p < 0.05), while much lower values Surface 1.84 ± 0.01 171 ± 2 0.8 ± 0.3 Medium 1.7 ± 0.2 20 ± 2 were observed for surface and ambient Burrow wall 1.44 ± 0.08 147 ± 4 19.2 ± 0.50 Poor 6.3 ± 1.0 25 ± 2 –1 Ambient 1.84 ± 0.02 173 ± 4 1.2 ± 0.1 Medium 1.6 ± 0.1 30 ± 4 sediment (0.85 and 0.60 µg g respec- tively). Papaspyrou et al.: Properties of Pestarella tyrrhena burrows 185

Table 2. Spearman-rank correlation analysis among variables analysed. TOC = (3.6%), showing a maximum of 7.9% in total organic carbon, PRT = protein, CHO = carbohydrate, LPD = lipid, Chl a = August 2002, while burrow wall and chlorophyll a, Phaeo = phaeopigments. Number of observations: n = 12 with ambient sediment had a mean of 0.7 to the exception of bacteria where n = 6. Significant values reported as: *p < 0.05, **p < 0.01 0.8% (Fig. 3).

TOC PRT CHO LPD Chl a Phaeo Bacteria Biochemical composition of organic TOC – matter PRT 0.699* – CHO 0.762** 0.755** – The 3 sediment types always had sig- LPD 0.678* 0.545 0.895** – Chl a 0.398 0.237 0.594* 0.727** – nificant differences in protein, carbo- Phaeo 0.524 0.615* 0.846** 0.902** 0.790** – hydrate, and lipid content, with yearly Bacteria 0.828* 0.942** 0.942** 0.657 0.314 0.600 – mean being highest at the burrow wall (p < 0.05) (Fig. 4). Sediment protein con- tent in the burrow wall was 4.2 and The highest yearly mean ratio of chl a to CPE, indicat- 6.5 times higher than ambient sediment and surface sed- ing fresher material, was found for surface sediment iment respectively, while for carbohydrates a 11-fold in- (64%), with a maximum in August (87%) (Fig. 2). Bur- crease was observed at the burrow wall compared to the row walls, on the other hand, showed the lowest ratios 2 other sediment types. Burrow wall lipid content was 4.0 (33%), while ambient sediment was intermediate. Over- and 4.9 times higher than surface and ambient sediment all, microphytobenthic carbon accounted for only a small respectively. The contribution of the various biopolymers fraction of TOC in all sediment types. Surface to TOC was different in the 3 sediment types. Carbohy- sediment had the highest yearly mean contribution drates were the most abundant type in the burrows (25%), followed by proteins (19%) and lipids (3%), while proteins were most important both in the surface (18%) and the ambient (25%), whereas lipids made only a small contribution (Fig. 3). Thus, the total identified frac- tion (PRT+ CHO+LPD) of TOC was highest in the bur- row wall (47%), followed by ambient sediment (40%) and lowest for surface sediment (36%) (Fig. 3).

Particulate iron

The total extracted particulate iron pool showed 2-fold and 3-fold higher concentration in the burrow

Fig. 3. Yearly mean relative contribution (%) of proteins (PRT), Fig. 2. (a) Seasonal variation in chlorophyll a content carbohydrates (CHO), lipids (LPD) and microphytobenthic (µg [g dw]–1) and (b) chlorophyll a to total chloroplastic pig- chlorophyll a (MPB) carbon to total organic carbon (TOC) of ments (CPE) (%) of surface sediment, burrow wall, and ambi- surface sediment, burrow wall, and ambient sediment asso- ent sediment associated with burrows of Pestarella tyrrhena ciated with burrows of Pestarella tyrrhena (mean + SE, n = 4) 186 Aquat Microb Ecol 38: 181–190, 2005

Fig. 4. Yearly mean protein (PRT), carbohydrate (CHO) and Fig. 6. Bacterial abundance in surface sediment, burrow wall, lipid (LPD) content (mg [g dw]–1) of surface sediment, burrow and ambient sediment associated with burrows of Pestarella wall, and ambient sediment associated with burrows of tyrrhena (mean + SE, n = 3) Pestarella tyrrhena (mean + SE, n = 4) wall compared to the ambient and surface sediment Bacterial diversity respectively (p < 0.05) (Fig. 5). Fe(III) constituted 62% of the total iron pool in the burrows, 33% in the surface Visual inspection of DGGE profiles showed marked sediment, while only small amounts of Fe(III) were changes in the number and position of the bands, found in the ambient sediment (5%). suggesting seasonal changes and effects of location (Fig. 7). Cluster analysis of the band patterns showed 2 distinct groupings; the anoxic ambient sediment sam- Bacterial abundance ples showing the highest similarity (65%) and the bur- row wall sediment samples grouped together with a Mean bacterial abundances based on DAPI counts similarity of 45%. These 2 groups joined together at a were 1 order of magnitude higher in the burrow wall similarity of 30%. Surface sediment samples were (4.0 × 109 cells cm–3) than in both surface (4.5 × 108 cells more variable, suggesting significant bacterial com- cm–3) and ambient sediment (4.8 × 108 cells cm–3) (p < munity changes during the course of the year at the 0.01) (Fig. 6). Bacterial abundance was affected by the sediment-water(air) interface. season, being higher in summer than in winter; how- ever, this was statistically confirmed only for surface and burrow wall sediment (p < 0.05). DISCUSSION

Burrow wall characteristics

Pestarella tyrrhena burrows in Vravrona Bay are plastered with a smooth red/brown layer, consisting of compacted sediment with a much finer grain composi- tion than ambient sediment, as has been reported pre- viously from other thalassinidean burrows (Dobbs & Guckert 1988, de Vaugelas & Buscail 1990, Over 1990). This suggests that P. tyrrhena is capable of sorting sedi- ment grains; selected fine particles are incorporated into the burrow wall to stabilise the structure, resulting in a lower grain- size homogeneity along burrow walls. Even though particle size selectivity by P. tyrrhena has not been studied extensively, such a mechanism during feeding is known for other thalassinidean shrimps with Fig. 5. Particulate Fe pools in surface sediment, burrow wall, a similar mode of life (Pinn et al. 1998, Stamhuis et al. and ambient sediment associated with burrows of Pestarella 1998). Fine particles may also be transported from the tyrrhena in January 2002 (mean + SE, n = 3 to 4) overlying water into the burrow during irrigation, Papaspyrou et al.: Properties of Pestarella tyrrhena burrows 187

fusing horizontally from ambient anoxic sediment and subsequent precipitation in the oxic zone of the burrow wall (Aller & Aller 1986, Over 1990).

Sources of burrow wall organic matter

Organic matter derived from microalgae (i.e. phyto- plankton and microphytobenthos) may constitute the main food source for deposit feeders in coastal envi- ronments (MacIntyre et al. 1996). The content of chl a in the burrows of Pestarella tyrrhena was similar to or slightly higher than that in surface sediment, as previ- ously reported for other burrowing shrimps (Dobbs & Guckert 1988, Kinoshita et al. 2003). One source of chl a in the burrows could be phytoplankton from the overlying water or resuspended microphytobenthic biomass translocated into the burrows by irrigation. Assuming an irrigation rate estimate of 5.8 l d–1 ind.–1 (Papaspyrou et al. 2004), a mean annual water column Fig. 7. Clustering by unweighted pair group means with chl a concentration of 4.8 µg l–1 at Vravrona Bay (S. arithmetic averages (UPGMA) of the band patterns obtained by denaturant gradient gel electrophoresis (DGGE) analysis Papaspyrou et al. unpubl. data) and ignoring degrada- of samples taken around Pestarella tyrrhena burrows. The tion, a maximum of 53% of the chl a in the burrow wall amount of amplified product obtained for surface sediment could be attributed to daily transport of microalgae from April 2002 was too small to detect the individual bands. from the overlying water if all chl a is retained during Major bands were identified manually from densitometric traces (not shown). The band patterns shown are generated water passage. Another source of chl a in the burrows from the original traces after background subtraction could be microphytobenthos buried into the sediment due to sediment reworking, which constantly buries and re-exposes diatoms at the surface. A high ben- where they stick to the burrow wall (Dobbs & Guckert thic microalgal biomass (mainly diatoms) has been ob- 1988, Kristensen 1988). served in the presence of bioturbating infauna even at A number of previous investigations have not detected a depth of 25 cm (Branch & Pringle 1987, Andresen & any difference between organic matter in burrow walls Kristensen 2002, Katrak & Bird 2003). However, the of thalassinidean shrimps and surrounding sediment overall low contribution (Fig. 4) and the non-significant (Dworschak 1983, Bird et al. 2000) while others, in accor- relationship of autotrophic carbon to TOC (Table 2) for dance with our results, have found significant enrich- the sediment of Vravrona Bay, indicates that the domi- ments (de Vaugelas & Buscail 1990, Dworschak 2001). nant organic matter sources are other than microalgae. Based on Pestarella tyrrhena burrow dimensions deter- Infaunal mucous secretions have been suggested as mined from in situ studies (Panaritou 1995, Dworschak a source of organic enrichment of burrow walls (Aller 2001, P. C. Dworschak pers. comm.) and our TOC mea- et al. 1983, Kristensen et al. 1985). Dworschak (1983, surements, and assuming an equal distribution of or- 1998) suggested that Pestarella tyrrhena lines vertical ganic carbon along the burrow walls (1 mm thickness), sections and roofs of the burrow with a mixture of fine- we estimate that for the mean population density at grained material and secreted mucus. In general, the Vravrona Bay (63 ind. m–2; Thessalou-Legaki 1987) be- presence or absence of a lining in thalassinidean tween 2 and 10% of the total sedimentary TOC in a shrimp burrows is determined to a large extent by given area (depending on burrow length) is located the habitat and sediment characteristics (i.e. sediment along the burrow walls. grain size). Thus, in sandy sediments, a mucus lining is Higher concentrations of trace metals along burrow necessary to support the easily collapsing walls of the walls, such as we observed here for iron, have been burrow, whereas this is not usually the case in cohesive reported previously (Aller et al. 1983, Aller & Aller muddy sediments (for a review see Dworschak 1983). 1986, Over 1990, Gribsholt et al. 2003). This increase Organic enrichment of the burrow wall may also be may be attributed to a higher proportion of clay parti- the result of seagrass detritus accumulation along the cles, allowing for sorption of iron oxides and/or organ- burrow wall. Posidonia oceanica detritus, transported ics (Aller & Aller 1986, Over 1990). In addition, the from adjacent subtidal areas, are abundant on the elevated iron concentration may result from Fe(II) dif- intertidal flat of Vravrona Bay. The detritus drifts by 188 Aquat Microb Ecol 38: 181–190, 2005

wave and current action and occasionally, creates Bacterial abundance and community characteristics large wrack beds on the shore. Drifting seagrass debris is also trapped in the funnel-shaped burrow openings As for other thalassinidean shrimp burrows (Aller et and may fall passively into the galleries or become al. 1983, Branch & Pringle 1987, Kinoshita et al. 2003), buried at the surface under sediment ejected from bacterial abundance within burrow walls of Pestarella the mound opening (Dworschak 2001, Papaspyrou et tyrrhena was significantly higher than at the sediment al. 2004). This material, together with older detritus surface and in adjacent ambient sediment. This is in already deposited in the sediment and selected by the agreement with Dworschak (2001), who, in addition, shrimp, can be seen incorporated into the burrow wall found an even higher abundance of bacteria in cham- or in burrow chambers, where it is allowed to decom- bers with seagrass debris inside P. tyrrhena burrows pose creating organic hot spots. than along burrow walls. These results are indicative Our values for biopolymeric compounds in the sedi- of a gardening activity. However, little is actually ment, particularly in the burrow walls, were among the known on what exactly is consumed and utilised for highest reported in the literature (Pusceddu et al. 1999, the animal nutrition during ‘wall-grazing’ by tha- their Table 2). However, the generally low chl a to CPE lassinidean shrimps like P. tyrrhena (Griffis & and protein to carbohydrate ratio (PRT:CHO = 0.6), Suchanek 1991). In addition, the role of gardening and as well as the dominance of carbohydrates for the bur- the significance of bacteria as a food source for deposit row wall of Pestarella tyrrhena at Vravrona Bay, are feeding organisms in general are still a matter for dis- characteristic of highly oligotrophic or detritic environ- cussion, since bacteria usually represent only a small ments (Danovaro et al. 1994, Danovaro 1996, Pusceddu fraction of the available organic carbon and nitrogen in et al. 1999) and indicate the presence of large amounts of sediments (Cammen 1980, Danovaro et al. 1994, aged detritus of relatively low nutritional quality derived Danovaro 1996, Andresen & Kristensen 2002). from e.g. seagrasses. Dobbs & Guckert (1988) also found The observed differences in the composition and a low chl a to CPE ratio in the burrow wall of Callianassa quality of organic matter (i.e. food availability), as well trilobata, which was attributed to the incorporation of as chemical character between burrow walls and sur- vascular plant debris into the burrow lining. rounding sediment were expected to affect not only Only a small fraction of Posidonia oceanica detritus is the abundance but also the composition of the bacter- directly available to benthic consumers for consump- ial communities. Previous studies based on phospho- tion, due to a low nitrogen content and a high content lipid fatty acid compositions have found no or only of refractory compounds, such as structural carbo- limited differences in bacterial composition between hydrates (Velimirov 1987, Danovaro 1996, Pusceddu et burrow wall and surrounding sediment for thalassi- al. 1999). The rapid deposition combined with the low nidean shrimps (Dobbs & Guckert 1988, Bird et al. nutritional quality of incorporated detritus suggests 2000). Our results based on more sensitive DGGE pro- that the sediment, and the burrow environment in files, on the other hand, show a clear distinction be- particular, act as an organic matter sink (‘detrital trap’) tween molecular fingerprints of bacterial communities (de Vaugelas & Buscail 1990, Pusceddu et al. 1999, from the different parts of Pestarella tyrrhena burrows. Kinoshita et al. 2003) where the fraction of primary Burrow wall and ambient sediment samples always production not directly available to the detritus food- form 2 separate groups, indicating that seasonal differ- chain accumulates. It therefore appears that organic ences are less important than location in determining nitrogen (i.e. protein) is the major limiting element for the bacterial community composition. In contrast, there deposit feeders (Tenore et al. 1984), and that they need were large and distinct seasonal changes in the com- to employ other mechanisms to obtain the needed munity of surface sediment from Vravrona Bay, reflect- nitrogen. Gardening of bacteria has been proposed as ing large variations in environmental conditions at a way to overcome this limitation. Many deposit feed- the sediment-water(air) interface (S. Papaspyrou et al. ing thalassinidean shrimps, like Pestarella tyrrhena, unpubl. data). Similarly, Lucas et al. (2003) reported store seagrass detritus in burrow chambers and that the bacterial composition along the burrow walls burrow walls to decompose (Griffis & Suchanek 1991, of Nereis diversicolor is stable over time and markedly Ziebis et al. 1996, Dworschak 2001), most likely in different from the community at the sediment surface. order to feed on the nitrogen-rich microorganisms Even though microorganisms along burrow walls have (bacteria and meiofauna) growing on this material. to adapt to the rapidly oscillating redox environment These typically show ‘wall-grazing’ behav- due to the usually intermittent burrow ventilation iour, where sediment is removed from the burrow wall, (Kristensen 1988, Ziebis et al. 1996, Furukawa 2001), sorted and part of it (including the microorganisms) our results confirm the idea that infaunal burrows, and ingested, while the remaining is either ejected from the ambient sediment, are stable environments over long- burrow or incorporated again into the wall. term periods compared to surface sediment, which is Papaspyrou et al.: Properties of Pestarella tyrrhena burrows 189

continually disturbed by macrofaunal feeding, and microbial properties of burrows of the deposit-feeding current or wave mediated action (Kristensen 1988). In Thalassinidean ghost shrimp Biffarius arenosus (Deca- addition, our results indicate that the bacterial commu- poda: ). Estuar Coast Shelf Sci 51:279–291 Bligh EG, Dyer W (1959) A rapid method for total lipid extrac- nity along burrow walls resembles that of the ambient tion and purification. Can J Biochem Phys 37:911–917 anoxic sediment more than that of the sediment sur- Branch GM, Pringle A (1987) The impact of the sand prawn face. The same has been observated for burrows of Callianassa kraussi Stebbing on sediment turnover and on N. diversicolor and Nereis virens (S. Papaspyrou et bacteria, meiofauna, and benthic microflora. J Exp Mar Biol Ecol 107:219–235 al. unpubl. data). This supports previous suggestions Buchanan JB (1984) Sediment analysis. In: Holme NA, Mc- based on physicochemical properties of burrows that Intyre AD (eds) Methods for the study of marine benthos. (1) burrow walls should not be considered a simple Blackwell Scientific Publications, Oxford, p 45–61 extension of the oxic sediment surface and (2) burrow Cammen LM (1980) The significance of microbial carbon in wall community structure should be significantly dif- the nutrition of the deposit feeding polychaete Nereis suc- cinea. Mar Biol 61:9–20 ferent regarding both the microbial ecology and dis- Danovaro R (1996) Detritus-bacteria-meiofauna in a seagrass tribution of functional groups (Kristensen et al. 1985, bed (Posidonia oceanica) of the NW Mediterranean. Mar Kristensen 1988, Furukawa 2001). Biol 127:1–13 Infaunal burrows are dynamic systems with intense Danovaro R, Fabiano M, Boyer M (1994) Seasonal changes of benthic bacteria in a seagrass bed (Posidonia oceanica) of small-scale gradients that support the growth of the Ligurian Sea in relation to origin, composition and fate complex microbial assemblages. The physicochemical of the sediment organic matter. Mar Biol 119:489–500 properties of the burrows, their age and the irrigation De Jonge VE (1980) Fluctuations in the organic carbon to pattern are amongst the factors that have been shown chlorophyll a ratios for estuarine benthic diatom popula- to affect the composition of the bacterial community tions. Mar Ecol Prog Ser 2:345–353 de Vaugelas J, Buscail R (1990) Organic matter distribution in along the burrow walls in artificial experimental sys- burrows of the thalassinid Callichurus laure, tems (Marinelli et al. 2002). Further studies are needed Gulf of Aqaba (Red Sea). Hydrobiologia 207:269–277 to clarify the mechanisms by which Pestarella tyrrhena Dobbs FC, Guckert JB (1988) Callianassa trilobata (Crus- bioturbation affects the bacterial community composi- tacea: Thalassinidea) influences abundance of meiofauna and biomass, composition, and physiologic state of micro- tion, which bacterial functional groups are favoured by bial communities within its burrow. Mar Ecol Prog Ser the burrow environment, and the possible role of bac- 45:69–79 teria in the recycling of aged detritus and P. tyrrhena Dworschak PC (1983) The biology of Upogebia pussila nutrition. (Petagna) (Decapoda, Thalassinidea). I.: the burrows. PSZN I: Mar Ecol 4:19–43 Dworschak PC (1987) Feeding behaviour of Upogebia pusilla and Callianassa tyrrhena (Crustacea, Decapoda, Thalas- Acknowledgements. We wish to thank all the friends and col- sinidea). Investig Pesq 51:421–429 leagues for their help during sample collection. Dr. I. Akou- Dworschak PC (1998) The role of tegumental glands in bur- mianaki is acknowledged for priceless advice during the row construction by two mediterranean Callianassid biopolymers analyses. We also acknowledge the valuable shrimps. Senckenb Marit 28:143–149 comments from 3 anonymous reviewers that helped us Dworschak PC (2001) The burrows of Callianassa tyrrhena improve the manuscript. This work was supported by grants (Petagna 1792) (Decapoda: Thalassinidea). PSZN I: Mar to M.T.-L. by the University of Athens (SARG Grant Ecol 22:155–166 # 70/4/3276) and E.K. by the Danish Science Research Furukawa Y (2001) Biogeochemical consequences of macro- Foundation (Grant # 21020463). During this work, S.P. fauna burrow ventilation. Geochem Trans 11:83–91, DOI: received a PhD scholarship from the Greek State Scholarships 10.1039/b108381c Foundation. Gerchakov SM, Hatcher PG (1972) Improved technique for analysis of carbohydrates in sediments. Limnol Oceanogr LITERATURE CITED 17:938–943 Gribsholt B, Kostka JE, Kristensen E (2003) Impact of fiddler Aller JY, Aller RC (1986) Evidence for localized enhancement crabs and plant roots on sediment biogeochemistry in a of biological activity associated with tube and burrow Georgia saltmarsh. Mar Ecol Prog Ser 259:237–251 structures in deep-sea sediments at the HEBBLE site, Griffis RB, Suchanek TH (1991) A model of burrow architec- western North Atlantic. Deep-Sea Res 33:755–790 ture and trophic modes in thalassinidean shrimp (Deca- Aller RC, Yingst JY, Ullman WJ (1983) Comparative biogeo- poda: Thalassinidea). Mar Ecol Prog Ser 79:171–183 chemistry of water in intertidal Onuphis (Polychaeta) and Katrak G, Bird FL (2003) Comparative effects of the large bio- Upogebia (Crustacea) burrows: temporal patterns and turbators, Trypaea australiensis and Heloecius cordi- causes. J Mar Res 41:571–604 formis, on intertidal sediments of Western Port, Victoria, Andresen M, Kristensen E (2002) The importance of bacteria Australia. Mar Freshw Res 54:701–708 and microalgae in the diet of the deposit-feeding poly- Kinoshita K, Wada M, Kogure K, Furota T (2003) Mud shrimp chaete Arenicola marina. Ophelia 56:179–196 burrows as dynamic traps and processors of tidal-flat Ausubel FW, Brent R, Kingston RE, Moore DD, Seidman JG, materials. Mar Ecol Prog Ser 247:159–164 Struhl K, Smith JA (1995) Short protocols in molecular Kostka JE, Luther GW (1994) Partitioning and speciation biology, 3rd edn. John Wiley & Sons, New York of solid-phase iron in salt-marsh sediments. Geochim Bird FL, Boon PI, Nichols PD (2000) Physicochemical and Cosmochim Acta 58:1701–1710 190 Aquat Microb Ecol 38: 181–190, 2005

Kristensen E (1988) Benthic fauna and biogeochemical pro- (Decapoda: Thalassinidea) and sediment. MSc thesis, cesses in marine sediments: microbial activities and University of Athens, Athens (in Greek) fluxes. In: Blackburn TH, Sorensen J (eds) Nitrogen Papaspyrou S, Thessalou-Legaki M, Kristensen E (2004) cycling in coastal marine environments. John Wiley & Impact of Pestarella tyrrhena on benthic metabolism in Son, Chichester, p 275–299 sediment microcosms enriched with seagrass and macro- Kristensen E, Andersen FO (1987) Determination of organic algal detritus. Mar Ecol Prog Ser 281:165–179 carbon in marine sediments: a comparison of two CHN- Parsons TR, Maita Y, Lalli CM (1984) A manual of chemical analyser methods. J Exp Mar Biol Ecol 109:15–23 and biological methods for seawater analysis. Pergamon Kristensen E, Jensen MH, Andersen TK (1985) The impact of Press, Oxford polychaete (Nereis virens Sars) burrows on nitrification Pinn EH, Atkinson JA, Rogerson A (1998) Particle size selectiv- and nitrate reduction in estuarine sediments. J Exp Mar ity and resource partitioning in five species of Thalassinidea Biol Ecol 85:75–91 (Crustacea: Decapoda). Mar Ecol Prog Ser 169:243–250 Lovley DR, Phillips EJP (1987) Rapid assay for microbially Porter KG, Feig YS (1980) The use of DAPI for identifying and reducible ferric iron in aquatic sediments. Appl Environ counting aquatic microflora. Limnol Oceanogr 25:943–948 Microbiol 53:1536–1540 Pusceddu A, Sarà, Armeni M, Fabiano M, Mazzola A (1999) Lucas FS, Bertru G, Höfle MG (2003) Characterization of free- Seasonal and spatial changes in the sediment organic living and attached bacteria in sediments colonized by matter of a semi-enclosed marine system (W-Mediterranean Hediste diversicolor. Aquat Microb Ecol 32:165–174 Sea). Hydrobiologia 397:59–70 Marinelli RL, Lovell CR, Wakeham SG, Ringelberg DB, Reichardt W (1988) Impact of bioturbation by Arenicola marina White DC (2002) Experimental investigation of the con- on microbiological parameters in intertidal sediments. trol of bacterial community composition in macrofaunal Mar Ecol Prog Ser 44:149–158 burrows. Mar Ecol Prog Ser 235:1–13 Reise K (2002) Sediment mediated species interactions in Marsh JB, Weinstein WJ (1966) A simple charring method for coastal waters. J Sea Res 48:127–141 determination of lipids. J Lipid Res 7:574–576 Rice DL (1982) The detritus nitrogen problem: new observa- McIntyre HL, Geider RJ, Miller DC (1996) Microphytobenthos: tions and perspectives from organic geochemistry Mar the ecological role of the ‘secret garden’ of unvegetated, Ecol Prog Ser 9:153–162 shallow-water marine habitats. I. Distribution, abundance Schäfer H, Muyzer G (2001) Denaturing gradient gel electro- and primary production. Estuaries 19:186–201 phoresis in marine microbial ecology. Methods Microbiol Meadows A (1991) Burrows and burrowing animals: an over- 30:425-468 view. In: Meadows PS, Meadows A (eds) The environ- Stamhuis EJ, Videler JJ, de Wilde PAWJ (1998) Optimal mental impact of burrowing animals. Zoological Society of foraging in the thalassinidean shrimp Callianassa sub- London, Clarendon Press, Oxford, p 1–11 terranea. Improving food quality by grain size selection. Miller DN (2001) Evaluation of gel filtration resins for the J Exp Mar Biol Ecol 228:197–208 removal of PCR-inhibitory substances from soils and sedi- Sternberg SR (1983) Biomedical image processing. Computer ments. J Microbiol Methods 44:49–58 16:22–34 Muyzer G, de Waal E, Uitterlinden AG (1993) Profiling of Tenore KR, Hanson RB, McClein J, Maccubin AE, Hodson RE complex microbial populations by denaturing gradient (1984) Change in composition and nutritional value to a gel electrophoresis analysis of polymerase chain reaction- benthic deposit feeder of decompositing detritus pool. Bull amplified genes coding for 16s rDNA. Appl Environ Mar Sci 35:299–311 Microbiol 59:695–700 Thessalou-Legaki M (1987) Contribution to the study of eco- Nelson DW, Sommers LE (1975) A rapid and accurate method logy and biology of Callianassa tyrrhena (Petagna, 1792) for estimating organic carbon in soil. Proc Indiana Acad (Crustacea, Decapoda, Thalassinidea). PhD thesis, Uni- Sci 84:456–462 versity of Athens, Athens (in Greek) Over DJ (1990) Trace metals in burrow walls and sediments, Velimirov B (1987) Organic matter derived from a seagrass Georgia Bight, USA. Ichnos 1:31–41 meadow: origin, properties and quality of particles. Overmann J, Tuschak C (1997) Phylogeny and molecular PSZN I: Mar Ecol 8:143–173 fingerprinting of green sulfur bacteria. Arch Microbiol Ziebis W, Forster S, Huettel M, Jørgensen BB (1996) Complex 167:302–309 burrows of the mud shrimp Callianassa truncata and their Panaritou K (1995) Relations between Callianassa tyrrhena geochemical impact in the sea bed. Nature 382:619–622

Editorial responsibility: Søren Rysgaard, Submitted: September 29, 2004; Accepted: December 22, 2004 Silkeborg, Denmark Proofs received from author(s): February 8, 2005